August 21, 2020
Twirling introduces a single-chip low-power module for immersive audio conferencing
August 2020, Nuvoton and Twirling have launched a single-chip low-power voice module for immersive audio conferences, which allows everyone's voice to be captured clearly and accurately, while eliminating background noise.
An immersive voice module launched by Twirling supports microphone arrays for Ambisonic sound field capture. Ambisonic technology has witnessed fast adoption in immersive audio in recent years, not only in VR/AR, but also in the radio, film and television industry. Ambisonic audio can simultaneously supports the traditional mono, stereo transmission, as well as multichannel format such as 5.1,7.1.
The Twirling Immersive Audio Module features some state-of-the-art voice processing algorithms, including
- ● Soundfield capture and rendering
- ● Deep learning AI noise suppression capable of suppressing both stationary and non-stationary noise
- ● single channel or stereo echo cancellation
- ● sound source localization, beamforming
- ● Automatic Gain Control for Far and Near Field pickup
- ● Support various microphone pickup patterns
- ● Support USB interface
Based on deep learning (Deep Neural Network) techniques, Twirling AI noise reduction technology has the advantage of being able to track the ambient noise level quickly and accurately, while maintaining very high speech quality and minimizing distortion. It can effectively suppress both stationary and nonstationary noise, such as air conditioning noise, passing vehicles, office environment keyboard tapping sound. The algorithm has the characteristics of low delay and fast convergence, and meets the strict requirements of real-time voice communication. The algorithm can also improve the accuracy of speech recognition systems.
High performance, low power consumption
The voice module launched by Twirling adopts the Nuvoton ISD94100 series chips, the biggest feature is its low power consumption, which can effectively guarantee smooth and intelligent voice interaction capabilities, while reducing power consumption and product costs. The combination of low-power algorithms and chips means that the module can be better adapted to devices with batteries, and the application range can be extended to various small portable conference devices such as Bluetooth speakers and Bluetooth headsets, further enriching the category and landing of IoT The scene provides a more efficient and convenient user experience.
Features of ISD94100 chips:
- ● sadWith a Cortex®- M4 USB microcontroller, core computing speed up to 200 MHz, high capacity flash 512 K and SRAM192k
- ● PWM support 6 channels with 16 bit resolution
- ● Support USB FS 1.1 controller, compatible USB FS 2.0
- ● VAD support low power mode
- ● 4 digital MIC supported
This low-power voice module supports the free expansion of the module, can access Bluetooth, WiFi and other modules, and through USB and other interfaces to achieve a variety of data communication methods, not only for immersive meetings, intercom products, but also can cover such as smart home, smart car and other scenarios to meet the needs of different fields.
|Maximum noise suppression
||20 mm×20 mm
||USB I2C UART GPIO
Nuvoton focuses on the design and development of consumer ICs, voice ICs, multimedia ICs and other products, and has won the trust of customers with deep design capabilities. Attach importance to long-term relationships with customers and partners, and have R&D and service organizations in the United States, China, Israel and other places to strengthen regional customer support services and global logistics management. In terms of quality, through strict production process control and quality control operations, the effectiveness of yield improvement, supply chain management and customer satisfaction has been strengthened.
Twirling Technologies (Twirling In Time) is a pioneer in speech and audio processing. At Twirling, we develop hardware and software to enable intelligent voice interaction interface, immersive audio creation and distribution. By leveraging our deep understanding in AI, signal processing, and audio, we build products and solutions with a wide range of applications in consumer electronics, education, security, VR/AR, and many other areas.
Address :1606, Block A1, Long Term Tiandi Building ,18 Suzhou Street, Haidian District, Beijing
Address :710,7th Floor ,699-1 Xuanwu Avenue, Xuanwu District, Nanjing